testsuite: portable make_data_file helper; drop hard /dev/urandom dependency

symlink-dirlink-basis.test and chdir-symlink-race.test both
require a multi-kilobyte non-trivial-content source file for the
rsync delta algorithm to exercise.  Both used dd / head against
/dev/urandom directly, which fails on platforms that don't ship
/dev/urandom (e.g. HPE NonStop).  The dd error gets swallowed by
'2>/dev/null' and the test then fails with a misleading 'failed
to create test file' that hides the real cause.

Add make_data_file <path> <size> to testsuite/rsync.fns.  Prefers
/dev/urandom when readable (kernel-provided randomness, fast),
falling back to a deterministic awk LCG seeded from PID and a
POSIX cksum of the destination path.  Output is constrained to
printable ASCII (33..126) so the helper survives two awk-portability
quirks:

  - printf '%c', 0 terminates the string in some awks, emitting
    fewer than sz bytes;
  - gawk in UTF-8 locales encodes printf '%c', N for N > 127 as
    a 2-byte UTF-8 sequence, emitting more than sz bytes.

The tests don't need 8-bit binary entropy -- they just need
non-trivial bytes for rsync's block-matching algorithm.

Update both call sites to use the helper.  Linux/FreeBSD/macOS
still take the /dev/urandom fast path; NonStop and any other
platform missing the device get the awk fallback transparently.
Both paths verified locally with the symlink-dirlink-basis test.
This commit is contained in:
Andrew Tridgell
2026-05-21 07:28:03 +10:00
committed by Andrew Tridgell
parent e1c5f0e93a
commit 47e087d8eb
3 changed files with 51 additions and 3 deletions

View File

@@ -66,8 +66,10 @@ ln -s "$outside" "$mod/subdir"
# different content, mode 0666 (the perms the attacker tries to push).
SIZE=$(stat -c %s "$outside/target.txt" 2>/dev/null \
|| stat -f %z "$outside/target.txt")
head -c "$SIZE" /dev/urandom > "$src/target.txt"
head -c "$SIZE" /dev/urandom > "$src/subdir/target.txt"
make_data_file "$src/target.txt" "$SIZE" \
|| test_fail "failed to create source file"
make_data_file "$src/subdir/target.txt" "$SIZE" \
|| test_fail "failed to create source file"
chmod 0666 "$src/target.txt" "$src/subdir/target.txt"
cat > "$conf" <<EOF

View File

@@ -235,6 +235,50 @@ makepath() {
}
###########################
# Create a file at $1 of $2 bytes containing non-trivial content
# suitable for rsync's delta algorithm to chew on. Prefers
# /dev/urandom for speed and entropy, falling back to a
# deterministic awk pseudo-random generator on platforms that
# lack /dev/urandom (e.g. HPE NonStop). The tests using this
# helper don't need cryptographic randomness -- they only need
# bytes that compress and delta-match like normal file content.
make_data_file() {
if [ $# -ne 2 ]; then
echo "usage: make_data_file PATH SIZE" >&2
return 2
fi
if [ -r /dev/urandom ] && \
dd if=/dev/urandom of="$1" bs="$2" count=1 2>/dev/null && \
[ -s "$1" ]; then
return 0
fi
# Fallback: a 32-bit linear congruential generator with BSD/glibc
# parameters. Seeded from PID and a POSIX cksum of the destination
# path so successive calls with different paths produce distinct
# content. Output is constrained to the printable-ASCII range
# (33..126, i.e. '!' through '~') for two portability reasons:
# - awk implementations vary on whether printf "%c", 0 emits a
# NUL byte or terminates the string;
# - gawk in UTF-8 locales encodes printf "%c", N for N > 127
# as a 2-byte UTF-8 sequence, which would make the output
# larger than the requested sz.
# The tests using this helper don't need 8-bit binary data, only
# non-trivial content for the rsync delta algorithm.
_path_seed=$(printf '%s' "$1" | cksum 2>/dev/null | awk '{print $1}')
awk -v sz="$2" -v seed_a="$$" -v seed_b="${_path_seed:-0}" 'BEGIN {
s = (seed_a + seed_b) % 2147483648
if (s < 0) s = -s
for (i = 0; i < sz; i++) {
s = (s * 1103515245 + 12345) % 2147483648
b = (int(s / 65536) % 94) + 33 # 33..126
printf "%c", b
}
}' > "$1"
}
###########################
# Run a test (in '$1') then compare directories $2 and $3 to see if
# there are any difference. If there are, explain them.

View File

@@ -46,8 +46,10 @@ export RSYNC_RSH
# Helper: create a large file suitable for delta transfers.
# ~32KB is large enough for rsync's block matching to find matches.
# make_data_file lives in rsync.fns and falls back to an awk PRNG
# on platforms without /dev/urandom (e.g. HPE NonStop).
make_testfile() {
dd if=/dev/urandom of="$1" bs=1024 count=32 2>/dev/null \
make_data_file "$1" 32768 \
|| test_fail "failed to create test file $1"
}